Helicopter short-term collision avoidance with sampling-based model predictive control

Automatic helicopter flight in uncertain surroundings remains a challenging task due to sudden changes in environment, requiring fast response to guarantee safe and collision-free guidance. Increasing numbers of small unmanned aerial vehicles, which are not covered by air traffic control, pose a potential threat to rotorcraft operating in lower airspace. To provide collision avoidance in this scenario, the capability of reacting immediately to appearing obstacles and guiding the rotorcraft along feasible evasive trajectories is required. This paper presents an approach to short-term collision avoidance based on model predictive techniques. The proposed method, originally developed for automotive applications, finds optimal control inputs by predicting a set of trajectories utilizing a model resembling the helicopter dynamics. Compared to model predictive control, no iterative optimization is adopted, resulting in deterministic execution time. The proposed method is evaluated by closed-loop simulations with a non-linear helicopter model. Additional hardware-in-the-loop simulations are conducted to examine the real-time capability of the approach.